The integration of ‘AI’ technologies into weapon systems introduces a complex dimension to international relations and security, championing technological solutions for enduring warfare challenges, notably enhancing ‘situational awareness’ through advances such as automated ‘vision’. However, the discourse, particularly in Western militaries like that of the United States, often overlooks inherent limitations and issues in AI-based warfare. This paper explores ‘AI’s’ implications for military vision by inter alia scrutinising the US military’s Joint All-Domain Command and Control (JADC2) process. It argues that the US military actively transforms the observation, decision, and action apparatus, progressively substituting human vision and decision-making, leading to a multidimensional de-visualisation. This denotes fundamental changes in human perception, reshaping knowledge, control, and agency dynamics. In conclusion, the paper suggests an imminent era of de-visualisation in the military – a deliberate relinquishment of human control for perceived military efficiency and effectiveness. This marks a transformative shift, urging nuanced consideration of the profound impact of ‘AI’ technologies on warfare dynamics.